An Ensemble Two-Level Classification Model (Etlcm) for Students’ Performance Prediction and Classification

Abstract

Educational data mining (EDM) is a type of learning analysis that employs machine learning
to classify academic data sets at various levels. Knowledge Tracing and student performance
prediction are Knowledge Tracing (KT). Student performance prediction is a typical activity
in EDM, with the goal of identifying poor or at-risk students. Early identification of these
students allows for the implementation of supporting measures that may aid in their
improvement. Furthermore, it assists academicians in taking the essential measures and
stages to train pupils in order to improve their performance. The proposed methodology in
this research effort intends to produce an evolutionary based students' academic performance
prediction, which aids in both improving their weaknesses and managerial decision-making.
Because the obtained data initially contained useless and redundant information, the raw
dataset was pre-processed by deleting the irrelevant and missing values. Faculty, students,
and technology employed are the three key characteristics that are used to make feature
selections. Students' academic performances are predicted using the selected features by
building an Ensemble Two-Level Classification Model (ETLCM). The prediction model
significantly enhances the accuracy of student classification based on personal and academic
criteria. The simulation result clearly demonstrates that the performance of this research
methodology generates more precise results than existing EDM approaches.